Cross-region replication

Last Updated: Aug 15, 2017

Bucket Cross-Region Replication automatically and asynchronously copies objects in buckets across different OSS data centers. It also synchronizes changes to objects in the source bucket and copies them across to the target bucket, enabling support for cross-region disaster recovery of buckets.

Application scenarios

Scenarios in which enabling Bucket Cross-Region Replication is recommended are as follows:

  • Compliance requirements
    Through cross-region synchronization, data can be copied between OSS data centers at different locations to satisfy compliance requirements such as specified distances between objects.

  • Latency
    In order to minimize object access latency between objects at different geographical points, a copy of the desired object can be maintained at an OSS data center that is physically closer to users.

  • Data backup and disaster recovery
    In the event of natural disasters, backup data can be saved to another OSS data center to maintain data security and availability.

  • Atypical operations
    If you have computing clusters in different OSS data centers, and wish to use them to analyze the same group of objects, you can maintain copies of the desired objects copies at the different data centers.

Supported synchronization functions

Currently, cross-region synchronization supports buckets with different names. For two buckets in different regions, a user can sync the data in the source bucket to the target bucket using the following features:

  • Real-time data synchronization
    Monitor data addition, deletion, and modification in real time and sync changes to the target region bucket. For files of 2M or larger, synchronization may take several minutes. This ensures the ultimate consistency of the data on both sides.

  • Historical data migration
    Synchronize historical data in the source bucket, creating two identical data copies.

  • Real-time synchronization progress retrieval
    Display the latest synchronization time node for real-time data synchronization. For historical data migration, this feature shows the percentage of data migrated.

  • Easy configuration
    Use the management interfaces on the OSS Console to easily monitor and perform actions based on site requirements.

Restrictions

  • Currently, cross-region synchronization is available in Mainland China. Support for other regions will be announced when available.

  • The two buckets to be used for data synchronization must belong to different regions. Data synchronization cannot be performed between buckets in the same region.

  • Users can simultaneously operate on two buckets in the synchronization status. However, objects copied from the source bucket may overwrite any objects of the same name in the target bucket. Use this function with caution.

  • Because Bucket Replication uses an asynchronous copying method, it may take time for data to be fully copied to the target bucket. Prepare accordingly to ensure the process can be completed properly.

  • Cross-region synchronization applies only when the two buckets to be synced do not enable data to be synced to or from a third bucket. For example, if synchronization is activated from Bucket A to Bucket B, the user cannot activate synchronization from Bucket A to Bucket C before deleting the synchronization configuration between Bucket A and Bucket B.

Reference

Thank you! We've received your feedback.